September 2013 43 Reducing Recidivism Through Probation Supervision: What We Know and Don’t Know From Four Decades of Research

Chris Trotter Monash University, Australia

THIS ARTICLE IS about the relationship use in criminal justice interventions and how recidivism in community-based supervision, between recidivism rates and supervision they relate to recidivism. They found the fol- few if any meta-analyses or other literature skills used by probation officers (or others lowing core practices to be significantly related reviews have focused on this issue. This review who supervise offenders on community-based to reduced recidivism: relationship factors, is an attempt to fill this gap. orders or parole). It focuses on routine day- skill factors, effective reinforcement, effec- This article takes the form of a literature to-day supervision rather than on intensive tive disapproval, problem solving, structured review rather than a meta-analysis. While supervision programs or other specialized learning, effective modelling, and effective meta-analysis has done much to further knowl- programs or interventions. use of authority. The meta-analysis, however, edge about effective practice in criminal justice, The general issue of what works and what is not clear about the extent to which the it has also drawn some criticism. Berk (2007), doesn’t work with offenders has received a studies included in the meta-analysis focused for example, argues that because meta-analysis lot of attention since the Martinson report on routine community-based supervision of uses data that has not been generated through (1974) suggesting that nothing works. Many offenders rather than on more specialist or random sampling, the statistical conclusions meta-analyses have been undertaken, covering group interventions. are not valid. He recommends the use of many hundreds of studies. These meta-analyses There is some debate about the conclu- conventional methods of research reviews. have attempted to identify the characteristics sions reached in some of the meta-analyses Others have criticized meta-analysis for bias in of effective practices and in many cases have regarding the impact of different staff skills selection of studies towards those that have sig- attempted to quantify the impact of different and also on the emphasis placed by research- nificant results, for including studies with poor types of intervention (e.g., Andrews & Dowden, ers on various aspects of the skills. Fortune, methodology, and for combining different ways 2006; Lipsey & Cullen, 2007). Andrews and Ward, and Willis (2012), for example, argue of measuring recidivism (see Pratt, 2012, for a Dowden (2006), for example, argue based on that a focus on collaboration, offender goals, summary and rebuttal of criticisms). their meta-analysis that effective practice can offender strengths, and a sense of meaning Literature reviews may therefore also have be conceptualized as focusing on the principles are more important than the focus on risk and a place in developing knowledge about what of risk, needs, and responsivity. They suggest risk reduction that is supported by Andrews works and what doesn’t in criminal justice that effective practice concentrates on medium- and Bonta (2008). Fortune, Ward, and Willis settings. Jesson, Matheson, and Lacey have to high-risk offenders, criminogenic needs, and (2012) maintain that offenders are likely to argued that systematic literature reviews pro- the delivery of programs or interventions that desist from crime through a process of chang- vide a more rigorous approach to synthesizing take account of individual needs and make use ing identity, a process that involves movement the literature on a particular topic compared of structured cognitive behavioral techniques towards personal goals and community and to the more open style of conventional reviews or interventions. social support. It seems clear that there is (2011). According to Jesson and colleagues The meta-analyses undertaken to date no universal agreement about what works in (2011), systematic reviews are clear about have primarily concerned community-based offender programs or offender supervision their aims, what databases have been searched, interventions, including group programs and (see McNeill, Raynor, & Trotter, 2010, for and what studies have been included and specialized programs such as drug treatment more detailed discussion of this issue). excluded and why; in addition, systematic and clinical interventions. A meta-analysis There is also no universal agreement reviews have a narrow focus and report on the covering studies up to 1998 undertaken by regarding what works in routine offender quality of studies that have been examined. Dowden and Andrews (2004) attempted to supervision. While a number of studies, The review presented in this article is more identify core staff skills—in other words, the particularly in recent years, have examined akin to a systematic review than to a conven- specific practices that human service workers the relationship between staff practices and tional literature review (Jesson et al., 2011). 44 FEDERAL PROBATION Volume 77 Number 2

Research Question only five studies that examined individual practice of reflective listening was also related skills (e.g., relationship skills, problem-solving to recidivism when accompanied by appropri- The research question that I examine here is: skills) for their relationship to recidivism. In ate use of authority. What is the impact on offender recidivism of some cases only one or two skills or prac- Trotter (1996) did a study based on prin- different worker skills and practices used by tices were highlighted: for example the use ciples similar to those used in the Andrews supervisors in the one-to-one supervision of of single-case study (Vered Slonim-Nevo, et al. (1979) study. File notes were examined offenders on probation or other community- 1999), working with offenders on family issues in more than 300 adult probation and parole based orders? (Denning & Homel, 2008), or socialization client files (of more than 50 officers). Using Offender recidivism is defined in terms levels of workers (Trotter, 1990). These studies a regression analysis, the author found that of the measures used in the various stud- have not been included in this review. recidivism rates were significantly lower than ies, including rearrest, re-conviction, further Several studies have examined the impact those of a control group when workers showed offense, or failure to comply with conditions of of training on the performance of commu- evidence in file notes of use of prosocial mod- the court order. Most of the studies have used a nity corrections officers and the subsequent eling and problem solving, although problem two-year follow-up period for recidivism. impact on recidivism. While this review is solving only related to failure to comply Literature Search concerned with the impact of skills on recidi- with conditions. The study also examined vism rather than with the impact of training, role clarification and empathy, neither of I searched criminal justice abstracts and I have included these studies where they have which significantly related to low recidivism, ProQuest Criminal Justice (a comprehensive shown that the training has influenced the although role clarification was used more database of U.S. and international criminal performance of the probation officers and the often with high-risk offenders. justice journals) using the terms probation, study has considered the recidivism of clients Taxman (2007, 2008) examined a project effectiveness, recidivism, community, skills, supervised by the trained officers. that implemented (through training, supervi- and supervision. The search also used other sion, and management) an evidence-based methods recommended by Jesson et al. (2011), The Studies approach to supervision. A total of 274 adult including scanning reference lists of articles Eight studies have been identified and probation clients supervised by officers in the consulted, consulting with colleagues with included in this review consistent with the Practice Community Supervision model were expertise in the topic, and manual searching criteria referred to above. Each of the studies then compared to 274 matched probation- of relevant journals. examined the relationship between the use ers receiving routine probation supervision. The search focused on studies that exam- of evidence-based practices in probation and The Practice Community Supervision model ined routine supervision on probation or recidivism or the relationship between train- included use of the Level of Supervision other court orders rather than specialist ing, use of practices, and recidivism. Five of Inventory Revisited to assess risk and need interventions. It excluded group work with the studies examined the relationship between factors, case plan, referral, learning about trig- offenders or studies that examined only par- a range of individual skills used by probation gers to offending, incentives, and sanctions ticular groups of offenders, such as drug users officers and recidivism. The others considered and review. Use of the model as a whole was or sex offenders. It included studies focusing the impact of a general set of skills but did not related to recidivism, although relationships on adults and young people, although care examine individual skills for their impact on between individual skills and recidivism were is taken to distinguish between the two. It recidivism. The studies are summarized here not reported. should be noted, however, that in some stud- in the order in which they were published. Pearson et al. (2011) examined a program ies young people may be classified as under 16 Several of the studies have built on knowledge in the United Kingdom known as citizenship. while in other studies young people include from earlier studies. This was a structured probation supervision those up to 25. The earliest study located, conducted by program, based on “what works” principles, The studies were then analyzed in terms Andrews et al. (1979), was published as a that aimed to engage offenders in targeted of their methodology, particularly in terms report by the Canadian government rather interventions complying with the risk prin- of sample size, use of statistical tests of sig- than in a refereed journal; however, it is ciple and included training in motivational nificance, and use of regression analyses to included in this review because it was the first interviewing and prosocial modeling; the isolate the impact of various skills and allow study on the relationship between workers’ offender also worked through problem-solv- for offender risk levels. The search gave pref- skills and offender recidivism reported in the ing modules depending on the offender’s erence to published articles in peer-reviewed databases and it was the precursor to a num- particular risk and needs. An experimental journals, on the understanding that the peer- ber of subsequent studies. The study analyzed and control group was made up of about 7000 review process ensures at least some degree more than 200 audiotapes of worker/client offenders. Through use of regression analy- of methodological rigor. The studies were interviews (workers could be professional sis and other statistical techniques, authors then analyzed according to the different skills probation officers or volunteers) in proba- concluded that the program had an impact used by the workers in the study, the nature of tion in Canada and used regression analysis on recidivism. Some associations between those skills, and the relationship of the skills to examine the relationship between workers’ skills and recidivism were reported. Pearson to recidivism. practices and recidivism. The authors found et al. (2010) found that the skill of promot- One of the difficulties encountered in the following practices of probation officers to ing contact with other agencies was related doing the review was that in most cases a be significantly related to reduced recidivism: to recidivism and that low- to medium- group of skills is described, usually in terms appropriate use of authority, problem solving, and medium- to high-risk offenders gained of evidence-based practice, and those skills prosocial modelling, and reinforcement. The most benefit from the citizenship program. as a group are related to recidivism. I located September 2013 REDUCING RECIDIVISM THROUGH SUPERVISION 45

This program is somewhat different from from 272 clients, including those supervised reported that those supervised by workers the others referred to in this review, as it by the trained officers and those in a control rated as using more evidence-based practices involved worksheets, community contact, and group. The results were somewhat mixed, had a further offense rate after two years of specific modules rather than examining rou- with trained officers who used more skills 62 percent, compared to 81 percent for those tine supervision. It is included nevertheless generally doing better, but with varying results supervised by workers with low ratings on the because it does report on the implementation across the different locations. The authors skills. Robinson et al. (2011) reported a rear- of evidence-based practices across a large pro- referred to limitations, including the fact rest or failure on supervision rate of 34 percent bation sample. that the experimental group was selected by for moderate- to high-risk clients supervised Bonta et al. (2011) examined audiotapes departmental administrators and the officers by untrained staff, compared to 26 percent of interviews between 80 officers and 143 of selected offender participants. for moderate- to high-risk clients super- their adult clients and examined structuring Raynor, Ugwudike, and Vanstone (forth- vised by trained staff. Raynor, Ugwudike, and skills, relationship-building skills, behavioral coming) examined videotapes of 75 clients Vanstone (forthcoming) found that 26 percent techniques, and cognitive techniques and how supervised by 14 staff in the probation service were re-convicted after two years when super- these relate to recidivism. The authors used a of the channel island of Jersey. They coded vised by more skilled officers compared to 58 random design and regression analysis to con- each tape for overall use of skills. Using regres- percent supervised by less-skilled officers. trol for extraneous variables and found that sion analysis, they found that the workers Six of the eight studies examined in this the use of skills was related to low recidivism. with more skills had clients with significantly review show that clients supervised by workers Bonta et al. found that cognitive techniques lower re-offending rates. They also found with more skills have lower recidivism than cli- had the strongest impact on recidivism. that individual skills of verbal and non-verbal ents supervised by workers with less developed Robinson, VanBenschoten, Alexander, and communication, motivational interviewing, skills. The differences were between 20 percent Lowenkamp (2011) examined tapes of more and problem solving were significantly related and 55 percent. In relation to the other two than 700 interviews between adult probation to lower recidivism after two years. studies, Andrews et al. (1979) reported strong officers and their clients. They examined the correlations between the use of individual use of active listening, role clarification, use The Impact of Probation skills and client recidivism but did not report of authority, effective disapproval, effective Officer Skills an overall impact of the use of skills. The only reinforcement and punishment, problem solv- All of the studies that could be located which study that has not shown clear differences ing, and use of the cognitive model. Using a examined the practices of probation officers, between recidivism rates of those supervised multivariate analysis of the data, they showed whether through examining file notes or by workers using evidence-based practices that the clients of those using the model had audiotapes of interviews or by direct observa- and other clients is the Smith et al. (2012) significantly lower recidivism. They did not tion, have found that when probation officers study. While their results generally favored report on the relationship between individual use evidence-based practice skills their clients those who were trained in and used evidence- skills and recidivism. have lower recidivism. All but one of the stud- based practices, the results were mixed. The Trotter (2012) directly observed interviews ies showed a significant difference between researchers found that clients of high-fidelity in a juvenile justice setting in Australia. The the recidivism rates of those supervised by officers (those rated as using more skills) had interviews were also audiotaped. This was more skilled officers and recidivism rates of lower incarceration rates and arrests for new one of the few studies undertaken with young those supervised by less skilled officers. The crimes but had more technical violations. The people (up to the age of 20 years). One hun- extent of the differences varied and in some authors discuss limitations in the methodology dred and seventeen interviews were observed cases varied according to risk levels of the that might explain the results. and then coded for use of various skills, such clients (as I will discuss later). It seems reasonable to conclude that if as relationship, role clarification, prosocial Trotter (1996) reported that the clients of probation officers or others who supervise modeling, problem solving, and use of CBT those officers who showed evidence in file offenders on court orders use evidence-based techniques. The researchers used a global score notes of using the evidence-based model had a practice skills, their clients are likely to offend as a measure of overall use of the skills as well further offense rate after one year of 28 percent, less often. as scoring individual skills. The global score compared to 44 percent for those who did not was significantly related to client recidivism, use the model. After four years the difference Skills or Practices that Are after taking account of other factors through was 46 percent to 64 percent. Taxman (2007) Consistently Related to a regression analysis. The only individual reported a rearrest rate of 32 percent for cli- Reduced Recidivism items that reached or were close to statistical ents supervised in the Proactive Community While it seems clear that probation officers significance included the use of rewards and a Supervision group, compared to 41 percent with evidence-based practice skills are likely to non-blaming attitude by the worker. in the non-Proactive Community Supervision have clients with lower recidivism, there is less Smith, Schweitzer, Labrecque, and Latessa group. Pearson et al. (2010) reported a rate of clarity about the precise nature of the effec- (2012) in a United States study provided reconviction of 41 percent after two years for tive practice skills and which of those skills training to 21 youth and adult probation clients in the citizenship group, compared to contribute most to reductions in offending. In officers in effective practices including anti- 50 percent in the comparison group. Bonta et this section I outline the skills identified in the criminal modeling, reinforcement, effective al. (2011) reported a further offense rate of 25 studies and consider the extent to which each disapproval, structured learning, problem percent recidivism after two years for officers impacts recidivism. solving, cognitive restructuring, and relation- trained in effective practice skills, compared to ship skills. They then analyzed audiotapes 40 percent in a control group. Trotter (2012) 46 FEDERAL PROBATION Volume 77 Number 2

Prosocial Modeling and Reinforcement reinforcement is therefore a core component It seems that problem solving is a key Prosocial modeling and reinforcement has of each of the eight studies examined in this skill in effective supervision; however, there been included as one of the skills in each of review. It was significantly related to recidi- remains some doubt about the extent to which the studies, although in some of the studies vism in four of the five studies that considered problem solving should be a collaborative pro- it is defined as anti-criminal modelling and its direct relationship with recidivism. cess that involves working on offense-related issues as the client defines them or whether reinforcement (Bonta et al., 2011; Smith et Problem Solving al., 2012). Despite the different terminology, it should involve working on criminogenic prosocial modeling and reinforcement and Most of the studies refer to the use of prob- needs that emerge from a risk assessment anti-criminal modeling and reinforcement lem-solving techniques. Sometimes these are undertaken by the worker. included as part of cognitive techniques (e.g., share similar characteristics. Both involve The Use of Cognitive Techniques modeling prosocial values such as fairness, Bonta et al., 2011) and sometimes they are reliability, and non-criminal lifestyle and rein- defined as a separate skill (e.g., Trotter, 1996, Many of the studies have included the use of forcing statements and activities of offenders 2012). The definitions of problem solving are cognitive techniques. Perhaps the best example that reflect those values. They also involve nevertheless reasonably consistent across the of a cognitive technique is the ABC technique carefully and respectfully challenging pro- studies. The definitions commonly involve referred to by Lowenkamp, Alexander, and criminal comments and actions (e.g., making identifying offense-related problems (e.g., Robinson et al. (forthcoming), which involves excuses for offending). Taxman (2007:19) family issues, accommodation, drugs), setting teaching offenders about Antecedents that summarizes the concept in terms of “using goals to address the problems, and then devel- lead to Behaviors that lead to Consequences. incentives and sanctions to shape offender oping strategies to address the goals. There is, As already mentioned, there is some overlap behaviours” (p. 19). however, variation in the way problem solving in the various publications between the defi- Most of the studies examined in this review is undertaken, particularly in terms of the nitions of cognitive skills and problem solving that have considered the relationship between extent to which the problems to be worked on and prosocial modeling and reinforcement prosocial modeling and reinforcement have and the goals that are set are developed by the skills. Bonta et al. (2011), for example, refer found strong relationships to low recidivism. clients, the worker, or the two in collaboration. to a cognitive technique known as cognitive Andrews et al. (1979) found that differential Trotter (1996, 2012), for example, emphasizes restructuring as including reinforcement and reinforcement of probationers’ prosocial and working with client definitions of problems. problem solving. Similarly, prosocial model- anti-criminal expressions and the expression Robinson et al. (2011) suggest that the most ing in Trotter (2012) includes helping clients of prosocial sentiments was related to low important aspect of the skill is allowing the to reframe their pro-criminal comments into recidivism. Trotter (1996) found that evidence client to articulate the problem and the poten- prosocial ones and reinforcing client com- of prosocial modeling in file notes was more tial solution. Taxman (2007) refers to working ments that reflect an understanding of the closely related to reductions in recidivism with one criminogenic need and at the same relationship between thoughts and behaviors. than any other skill, and Trotter (2012) found time working with an interest of the client in Despite the confusion over definitions, the the use of rewards by youth probation officers order to motivate the client to commit to the studies generally support the use of cognitive to be related to recidivism (although not quite change process. Bonta et al. (2011), on the techniques. Cognitive techniques were part at statistically significant levels). Raynor et al. other hand, emphasize working with crimino- of the overall group of skills in each of the (forthcoming) found prosocial modeling sig- genic needs that are identified through a risk studies, with the exception of the two earlier nificantly related to low recidivism at one-year assessment undertaken by the worker. studies by Andrews et al. (1979) and Trotter and two-year follow-up (although it was only Support for problem solving in whatever (1996). Some of the studies examined the statistically significant after one year). form is provided by most of the studies. relationship between cognitive techniques and Bonta et al. (2011), on the other hand, Andrews et al. (1979) found that problem recidivism. Bonta et al. (2011) found that the found that behavioral skills, including effective solving with a concrete community focus was only intervention techniques predicting lower use of reinforcement and disapproval, were significantly related to recidivism. Trotter recidivism were cognitive techniques—how- not significantly related to low recidivism. (1996) found that problem solving (empha- ever, as I have noted, cognitive techniques were They found cognitive skills (discussed below) sizing a focus on client-defined problems broadly defined. Raynor et al. (forthcoming) to be the only skills related to recidivism, after and goals) was related to recidivism but found cognitive restructuring significantly taking risk into account. Their definition of only significantly related to compliance with related to reduced offending, but only at one- cognitive skills was, however, a broad one conditions rather than re-offending. Raynor year follow-up. Trotter (2012) did not find a and incorporated some of the micro-skills (forthcoming) found that problem solving significant association between worker use of that other researchers (e.g., Trotter, 1996, was significantly related to reduced offend- cognitive behavioral techniques and recidi- 2012) referred to as prosocial modeling and ing at both one-year and two-year follow-up. vism; however, the author noted that cognitive reinforcement—for example, helping offend- Bonta et al. (2011) found cognitive techniques behavioral techniques were used infrequently. to be significantly related to recidivism and ers re-frame pro-criminal expressions into Worker-Client Relationship prosocial ones. included problem solving as part of cognitive Each of the other studies considered in techniques. Smith et al. (2012), Robinson et Again, there are varying definitions of the this paper incorporated the concept of pro- al. (2011), Pearson et al. (2010), and Taxman concept of the worker-client relationship. social modeling and reinforcement. They did (2007) all included problem solving in their Andrews et al. (1979) used a psychological test not, however, examine its specific relation- repertoire of skills, although they did not of empathy and found that scores on an empa- ship with recidivism. Prosocial modeling and examine its specific relationship to recidivism. thy scale were unrelated to client recidivism (a September 2013 REDUCING RECIDIVISM THROUGH SUPERVISION 47 finding replicated by Trotter, 1996). Similarly, (e.g., Trotter, 1996, 2012; Bonta et al., 2011) included in cognitive techniques by some the practice of reflective listening when iden- the Level of Supervision Inventory (Andrews researchers. Problem solving is focused on tified through the examination of audiotapes & Bonta, 2008) in one of its forms. client definitions of problems in some stud- of interviews was found to be unrelated to Smith et al. (2012) found most impact with ies and on risk-related problems in others. recidivism by Andrews et al. (1979). Bonta et high-risk offenders; however, the impact was The client-worker relationship in particular is al. (2011) examined relationship skills, which not statistically significant overall. Robinson defined in different ways in different studies. included role clarification and active listening et al. (2011), on the other hand, found among For example, different studies have examined skills; however, these were not found to be a group trained in effective practices that relationship in terms of empathy, active lis- independently related to reduced recidivism at the greater impact on offending was with tening, trusting, non-blaming, expectation, statistical levels. Trotter (forthcoming) found moderate-risk offenders, with less impact engagement in interviews, ground rules, and that clients who were judged by the observer on high-risk offenders. Trotter (1996, 2012, role clarification. to be disengaged in the interview still ben- forthcoming) found in both studies that low- Nevertheless, all of the studies that have efited from the use of skills by their workers. medium and high-risk offenders all had lower examined the use of skills and client recidi- Engagement in the interview was defined as a re-offending when their workers had good vism found that when probation officers used relationship measure. skills. Pearson (2010) found that medium- certain practice skills, their clients offended On the other hand, Raynor et al. (forth- to-low- and medium-to-high-risk clients less often. In seven of the eight studies this was coming) found that verbal and non-verbal benefited from skilled intervention but that at statistically significant levels. communication were related to low recidivism high-risk offenders did not. Taxman (2007) The more recent studies have built on at one year and two years, but not the way the included risk assessment as one of the effective the earlier studies, particularly on the sem- interview was set up or the legitimate use of practice skills that led to improved outcomes, inal work undertaken by Andrews et al. authority, both of which were also defined in but did not specify the impact of the various (1979). Each of the studies examined included the study as relationship skills (as opposed to skills on offenders with different risk levels. prosocial modelling, problem solving, and structuring skills such as problem solving). Similarly, Raynor et al. (forthcoming) did not worker-client relationship, with the more Trotter (2012) found a non-blaming attitude specify the impact on offenders with different recent studies also including cognitive tech- by the probation officer to be significantly risk levels. niques and the role that client risk levels may related to low recidivism. Smith et al. (2012) play in the supervision process. found that offenders who perceived a trusting Other Factors It seems reasonable to conclude that pro- relationship with their supervisor were signifi- A number of other factors were referred social modeling and reinforcement, problem cantly less likely to be arrested for a new crime. to in the various studies but have not been solving, and cognitive techniques are core The relationship was included in each of examined often enough to reach any firm con- skills for reducing recidivism in probation the other studies, albeit with different defini- clusions about them. Taxman (2007) found supervision. These three skills have been tions. Robinson et al. (2011) referred to active that community support and referral were present in the studies in one form or another listening, role clarification, and feedback. related to lower recidivism. Pearson et al. (even though cognitive skills were less spe- Taxman (2007) referred to expectation and (2011) found that contact with other agencies cifically examined in the earlier studies), and ground rules. These studies did not, however, was related to reduced recidivism. Pearson et they have generally shown significant associa- consider the specific interaction between the al. (2010) and Raynor et al. (forthcoming) also tions with recidivism. client-worker relationship and recidivism. found that motivational interviewing training It is difficult to reach any firm conclusions Like prosocial modeling and problem solv- was related to reduced recidivism. about the impact of the worker-client relation- ing, the varying definitions of “relationship” Role clarification has also been included in ship. Trusting and non-blaming relationships make it difficult to generalize about its nature some of the studies (Bonta et al., 2011; Smith with good communication seemed to be more or its relationship to reduced recidivism. The et al., 2012; Trotter, 1996, 2012) as part of the effective than those characterized by reflective studies seem to support a hypothesis that repertoire of effective practice skills. Bonta listening practices or even engagement of the the practice of active listening or the use of also refers to discussions about conditions of client in the interview. empathy may not necessarily be important in probation as being related to higher recidi- Of the five studies that examined risk, probation supervision; however, a process that vism (Bonta et al., 2011). There is, however, four found that medium-risk probationers leads to a trusting and non-blaming relation- insufficient data on the direct relationship benefited from skilled supervision more than ship may be. to recidivism to reach any firm conclusions high-risk probationers did. In two of the about these practices. studies, low-risk offenders exposed to effec- Risk Levels of Clients tive practice skills also had lower recidivism. Andrews and Dowden (2006), among oth- Conclusion The previous research generally suggests that ers, have argued that medium- and high-risk The research on effective probation supervi- medium- to high-risk offenders benefit from offenders generally benefit more from inten- sion has come a long way since the seminal intensive interventions; however, the super- sive correctional interventions and that study by Andrews and colleagues in 1979. The vision offered in the studies identified in low-risk offenders benefit less from intensive studies have some limitations, particularly this paper was not necessarily intensive. The interventions. The studies considered in this in terms of the varying nature of the defini- studies generally examined practice skills in paper provide varying support for this prin- tions of skills used by different researchers. routine probation supervision; they seem ciple. Each of the studies have used some sort For example, elements of prosocial modeling to suggest that when supervisors have good of actuarial assessment of risk; in some cases and reinforcement and problem solving are 48 FEDERAL PROBATION Volume 77 Number 2 skills, their clients offend less often regardless References Pearson, D., McDougall, C., Kanaan, M., Bowles, A., & Torgerson, D. (2011). Reduc- of their levels of risk. Andrews, D. A., & Bonta, J. (2008). The psychol- ing criminal recidivism: Evaluation of ogy of criminal conduct. Cincinnati: Ander- Further Research Citizenship, an evidence-based probation son Publishing. supervision process. Journal of Experimental We know that the application of certain skills Andrews, D., & Dowden, C. (2006). Risk prin- Criminology, 7:73–102. in probation supervision is likely to lead ciple of case classification in correctional Pratt, T. (2012). Meta-analysis and its discon- treatment: A meta-analytic investigation. In- to lower recidivism when compared to the tents revisited. Journal of Offender Rehabili- ternational Journal of Offender Therapy and absence of those skills. We know that certain tation, 35(1), 23–41. Comparative Criminology, 50(1):88–100. key skills relate to low recidivism; however, Raynor, P., Ugwudike, P., & Vanstone, M. (forth- Andrews, D. A., Kiessling, J. J., Russell, R. J., & the definitions of these skills in some cases coming). The impact of skills in probation Grant, B. A. (1979). Volunteers and the one remain unclear. Detail about the practice of work: A reconviction study. Criminology to one supervision of adult probationers. good supervision is not generally available and Criminal Justice. Toronto: Ontario Ministry of Correctional Robinson, C., VanBenschoten, S., Alexander, in the refereed journal articles, presumably Services. M., & Lowenkamp, C. (2011). A random because of the space limitations. Some of the Berk, R. (2007). Statistical inference and meta- (almost) study of Staff Training Aimed at studies have attempted to describe the skills analysis. Journal of Experimental Criminol- Reducing Re-arrest: Reducing recidivism examined in their studies in detailed reports, ogy, 3:247–270. through intentional design. Federal Proba- for example, Lowenkamp et al. (forthcom- Bonta, J., Bourgon, G., Rugge,T., Scott, T.-L., tion, 75(2). Yessine, A.K., Gutierrez, L., & Li, J. (2011). ing); however, in most cases the workbooks, Smith, P., Schweitzer, M., Labrecque, R. M., & Community supervision: An experimental training packages, and examples of good and Latessa, E. J. (2012). Improving probation demonstration of training probation officers not-so-good skills are not available in the pub- officers’ supervision skills: An evaluation in evidence-based practice. Criminal Justice lic domain. Publication of the precise nature of of the EPICS model. Journal of Crime & and Behavior, 38(11), 1127–1148. the skills, including transcripts of interviews, Justice, 35(2), 89–199. Denning, R., & Homel, R. (2008). Predicting re- would help further training and implementa- Taxman, F. (2007, May/June). The role of com- cidivism in juvenile offenders on communi- munity supervision in addressing re-entry tion of practice skills. ty-based orders: The impact of risk factors from jails. American Jails. Many other skills have also been identi- and service delivery. Journal of Offender Taxman, F. (2008). No illusions: Offender and fied as evidence-based practices in probation Rehabilitation, 46(3–4), 189–215. organizational change in Maryland’s Pro- but have not been specifically examined in Dowden, C., & Andrews, D. A. (2004). The im- active Community Supervision Efforts. the research referred to in this paper. Taxman portance of staff practice in delivering effec- Criminology & Public Policy, 7(2), 275–302. tive correctional treatment: A meta-analytic and Sachwald (2010), for example, refer to 18 Taxman, F. S., & Sachwald, J. (2010). Managing review of the literature. International Journal evidence-based practices, including family chaos: Techniques to improve the quality of of Offender Therapy and Comparative Crimi- therapies, drug testing, and staff qualifica- supervision. In F. McNeill, P. Raynor, & C. nology, 48(2), 203–214. tions. The cause of evidence-based practice in Trotter, Offender supervision: New direc- Fortune, C. A., Ward, T., & Willis, G. (2012). probation would be furthered by research on tions in theory, research and practice. Willan, The rehabilitation of offenders: Reducing these and other practices. Cullompton, 172–192. risk and promoting better lives. Psychiatry, Trotter, C. (1990). Probation can work. A Most of the research (Pearson et al., 2011, Psychology, & Law, 19, 646–661. research study using volunteers. Australian excepted) has been undertaken with relatively Jesson, J., Matheson, L., & Lacey, F. (2011). Do- Journal of Social Work, 43(2), 13–18. small samples. More work on the imple- ing your literature review: Traditional and Trotter, C. (1994). Effective supervision of of- mentation of effective practices across whole systematic techniques. London: Sage. fenders. Ph.D. thesis. LaTrobe University, organizations, such as that done by Taxman Lipsey, M., & Cullen, F. (2007). The effectiveness Melbourne. of correctional rehabilitation: A review of and Sachwald (2010), might provide informa- Trotter, C. (1996). The impact of different super- systematic reviews. Annual Review of Law tion regarding the potential for consistent and vision practices in community corrections. and Social Science, 3, 297–320. widespread implementation of the effective Australian and New Zealand Journal of Lowenkamp, C., Alexander, M., & Robinson, practice skills. Criminology, 29(1), 29–46. C. (forthcoming). Using 20 minutes wisely: Overall, however, the studies pointing Trotter, C. (2012). Effective supervision of Community supervision officers as agents to low recidivism rates among probationers young offenders. Trends and Issues in Crimi- of change. In J. Smykla & M. Crow (Eds.), nal Justice, 448. supervised by skilled workers clearly suggest Offender re-entry: 21st century issues. Burl- Trotter, C. (forthcoming). Effective supervision that probation supervision can reduce recidi- ington, MA: Jones and Bartlett. of young offenders. In Offender Compliance. vism. Hopefully future research can tell us Martinson, R. (1974). What works? Questions London: Routledge. more about how this is done and how the effec- and answers about prison reform. The Pub- Vered, S.-N. (1999). Evaluating the effectiveness tive practices can be implemented more widely. lic Interest, 35:22–54. of juvenile probation officers: An Israeli McNeill, F., Raynor, P., & Trotter C. (2010). Of- study. Journal of Offender Rehabilitation, fender supervision: New directions in theory, 28(3–4), 77–90. research and practice. Willan, Cullompton.